174 BELL SYSTEM TECHNICAL JOURNAL 



Again there is a marked decrease in the resistance of granular 

 carbon contacts with increase in voltage which had not been satis- 

 factorily explained. This fact suggested amongst other possibilities 

 that the conduction process may involve the passage of electrons 

 across gaps of molecular dimensions in the manner of a cold point 

 discharge. Field gradients of sufficient magnitude to extract electrons 

 from a solid must exist in these gaps with only a fraction of a volt 

 across the contacts. If this is the main process by which current 

 passes between contacts, microphonic action might well be associated 

 with a variation of the gap dimensions under strain. 



Again recent work on the theoretical strength of solids had led to 

 experimental results showing that under certain conditions solids 

 may, without fracture, be subjected to strains greatly exceeding those 

 heretofore obtained. This suggested the possibility that the micro- 

 phonic effect of contacts might after all be associated with the straining 

 of small junctions welded under pressure and current. 



In view of the speculative nature of the situation it was clear that 

 a new experimental attack on the problem was necessary. We have 

 been making such an attack during the last few years and I now turn 

 attention to some of the experimental results and the main conclusions 

 to be drawn from them. 



Recent Experimental Work 

 Statement of the Problem 



Since the essential element in the carbon microphone is the so- 

 called "loose contact," the first and most fundamental step toward 

 the understanding of the physics of microphones is the solution of 

 the problem of the "loose contact" when in its sensitive or microphonic 

 state. 



Measurements on microphones such as the handset have enabled us 

 to specify pretty accurately the conditions under which any two 

 granules within the structure operate when the microphone is trans- 

 mitting speech or sound. 



In addition to the voltage, which is limited to one volt per contact, 

 these conditions may be stated briefly either in terms of contact 

 forces or in terms of movements between centres of granules. When 

 you realize how small these are — particularly the movements between 

 centres of granules — you will, I think, not be surprised that the solu- 

 tion of the problem of the "loose contact" has been so long delayed. 



For the condition of reasonably loud speech the diaphragm motion 

 is about 



1 X 10 =^ cm., 



